Plasticity of cone photoreceptors in adult zebrafish revealed by thyroid hormone exposure

Vertebrate color vision is predominantly mediated by the presence of multiple cone photoreceptor subtypes that are each maximally sensitive to different wavelengths of light. Thyroid hormone (TH) has been shown to be essential in the spatiotemporal patterning of cone subtypes in many species, including cone subtypes that express opsins that are encoded by tandemly replicated genes. TH has been shown to differentially regulate the tandemly replicated lws opsin genes in zebrafish, and exogenous treatments alter the expression levels of these genes in larvae and juveniles. In this study, we sought to determine whether gene expression in cone photoreceptors remains plastic to TH treatment in adults. We used a transgenic lws reporter line, multiplexed fluorescence hybridization chain reaction in situ hybridization, and qPCR to examine the extent to which cone gene expression can be altered by TH in adults. Our studies revealed that opsin gene expression, and the expression of other photoreceptor genes, remains plastic to TH treatment in adult zebrafish. In addition to retinal plasticity, exogenous TH treatment alters skin pigmentation patterns in adult zebrafish after 5 days. Taken together, our results show a remarkable level of TH-sensitive plasticity in the adult zebrafish.


T4 treatment alters topography of lws1 and lws2 reporter patterning in adult lws:PAC(H) transgenic fish
The lws:PAC(H) transgenic reports lws1 expression with GFP and lws2 expression with dsRedExpress ("RFP").This line [Tg(LWS1/GFP-LWS2/RFP-PAC(H))#430, (kj15Tg)] has been shown to recapitulate the characteristic pattern of lws1 and lws2 mRNA-expressing cones in the adult zebrafish, in which lws1 is expressed in the ventral and nasal periphery, with some expression in the dorsal periphery, and lws2 is expressed centrally and dorsally 34,40 .Further, the lws1 and lws2 reporters reproduce the response of the native transcripts to TH 14 .In larval zebrafish, native lws1 mRNA and the GFP reporter of lws1 in lws:PAC(H) show increases in size of their expression domains in response to 100 nM TH treatment while native lws2 mRNA and the RFP reporter of lws2 in lws:PAC(H) domains decrease 14 .Previous studies have shown that treatment of athyroid juvenile lws:PAC(H) zebrafish with T4 rescues GFP (lws1) expression, indicating that the reporter construct remains plastic to TH treatment at the juvenile stage, similar to the behavior of the native lws array 14 .As such, we reasoned it would be appropriate to investigate plasticity using the lws:PAC(H) reporter line.
Confocal imaging and subsequent expression domain area analysis showed that after 5 days of treatment with T4, the GFP (lws1) expression domain significantly expanded in comparison to controls (Fig. 1A-F).In control retinas, the GFP domain was found to include approximately 60% of the retina, while in T4 treated retinas, the GFP domain covered over 90% of the retina (Fig. 1I).In contrast, the RFP (lws2) domain remained similar in both groups (approximately 60%, Fig. 1J).This resulted in an increased region of interspersed and/or coexpressing GFP + and RFP + cones (Fig. 1K).Five-micron tissue sections (Fig. 1G-G″′,H-H″′) showed that in retinas from T4-treated fish, the majority of RFP expressing cells also expressed GFP (Fig. 1H″′; Supplementary Fig. S1).This coexpression could indicate that both members of the transgenic array were being transcribed or that only GFP was transcribed, but RFP (protein) had not yet been degraded.In order to distinguish between these possibilities, our next studies focused on monitoring endogenous mRNA.

Five days of T4 treatment induce widespread cone transcriptional changes
Confocal imaging of retinas that underwent hybridization chain reaction (HCR) in situ for lws1 and lws2 revealed that after 5 days of T4 treatment, the lws1 expression domain expanded to a similar extent seen in the lws:PAC(H) reporter transgenic (Fig. 2A-F; Supplementary Fig. S2).Interestingly, lws2 transcript was undetectable by HCR after 5 days of T4 treatment (Fig. 2E).Because our previous study identified the increase in lws1-expressing cones as the result of individual cones switching opsins 14 , and because our current results obtained from the lws:PAC(H) transgenics showed widespread coexpression of GFP reporting lws1 and RFP reporting lws2, we interpret that the expanded lws1 domain and lack of lws2 transcript was likely due to opsin switching.In other words, cone photoreceptors that previously expressed lws2 have switched to express solely lws1.These results also suggest that the GFP:RFP coexpression seen in the lws:PAC(H) fish (Fig. 1D) was likely due to slow degradation of the dsRedExpress reporter.Further, qPCR data (Fig. 2G) showed that after 5 days of T4 treatment, lws1 transcript abundance increased, while lws2 transcript abundance significantly decreased, in agreement with the HCR confocal images.We next analyzed other genes known to exhibit altered transcriptional abundance or expression domains after TH treatment; namely gngt2b, gngt2a 20 , rh2-1, and rh2-2 14 .We found that the abundance of gngt2b, encoding a gamma subunit of transducin associated with LWS2 cones and other cone subtypes in the central retina 26,41 , and www.nature.com/scientificreports/known to be downregulated by T3 treatment in larvae, decreased after 5 days of T4 treatment of adults (Fig. 2G) 20 .This finding suggests that cone photoreceptor genes other than those encoding opsins remain plastic to the effects of TH even in adulthood.We also analyzed gngt2a, which encodes a paralogous gamma subunit associated with LWS1 cones 20,41 , and which showed an altered expression domain following T3 treatment in larvae, although transcript abundance did not change 20 .In T4 treated adults, abundance of this transcript was not altered (Supplementary Fig. S3).The rh2 array is another set of tandemly replicated opsins in zebrafish 13 , and the members of this array are known to be affected by TH treatment in larval and juvenile fish 14,20 .We found that exogenous T4 treatment of the adults resulted in decreased abundance of both rh2-1 and rh2-2 transcripts (Fig. 2G).

T4 treatment induces widespread cone transcriptional changes in as little as 24 h
We then tested whether shorter treatments would also generate changes in transcription of cone genes.Interestingly, the results of the 24-h treatment were similar to those of the 5-day experiment, suggesting that gene expression in cone photoreceptors responds rapidly to changes in T4 levels (Fig. 3A-F; Supplementary Fig. S2).
We found that after 24 h of T4 treatment, the lws1 expression domain expanded to cover the entire retina and lws2 expression became undetectable by HCR (Fig. 3D-F).The 24-h qPCR results were also similar to the 5-day results, showing that the transcript abundance of lws1 increased after treatment while the abundance of lws2, gngt2b, rh2-1 and rh2-2 transcripts significantly decreased (Fig. 3G), and that of gngt2a did not change (Supplementary Fig. S3).Confocal imaging of retinas that underwent HCR in situ revealed that after 12 h of T4 treatment, the lws1 expression domain expanded to a similar extent seen in 5 day and 24-h treatment groups (Fig. 4A-F; Supplementary Fig. S2).In contrast, however, the lws2 expression domain remained similar to that of controls (Fig. 4B,E; Supplementary Fig. S2).We found that most of the cells in which lws2 mRNA was detected also showed lws1 expression (Fig. 4D).Because mRNA half-lives in vertebrates exhibit a wide range, from minutes to several hours, these results could indicate that many LWS cones actively transcribe both opsin genes after 12 h of T4 treatment, or that the cones have switched to express lws1 while lws2 mRNAs remain 42,43 .The specific half-lives of cone opsin mRNAs have not yet been determined, to our knowledge.Our qPCR results corroborate our in situ data, showing a significant increase in lws1 transcript abundance but no change in lws2 transcript abundance (Fig. 4G).Further, the transcript abundance of the other genes we investigated (gngt2b, gngt2a, rh2-1, rh2-2) also did not change (Fig. 4G; Supplementary Fig. S3).These results suggest that 12 h is not sufficient for all T4-mediated transcriptional changes to be made in adult retina.
Confocal imaging of adult retinas that underwent HCR in situ revealed that after 7 h of T4 treatment, the lws1 expression domain expanded to a similar extent seen in the previous treatment groups and the lws2 expression www.nature.com/scientificreports/domain remained similar to that of controls (Fig. 5A-F; Supplementary Fig. S2).We found that most of the cones in which lws2 mRNA was detected also showed lws1 expression (Fig. 5D).Our qPCR results corroborate these data, showing a significant increase in lws1 transcript abundance but no change in lws2 transcript abundance (Fig. 5G).Further, the transcript abundance of the other genes we investigated by qPCR (gngt2b, gngt2a, rh2-1, rh2-2) also did not change (Fig. 5G; Supplementary Fig. S3).These results suggest that 7 h is not sufficient to generate all T4 mediated changes in adult cone transcription.

Exogenous T4 treatment of adult zebrafish alters skin pigmentation
The adult zebrafish exhibits a characteristic pattern of dark stripes consisting of melanophores (darkly pigmented cells) and iridophores (iridescent cells) alternating with light interstripes containing xanthophores (yellow/ orange pigmented cells) and iridophores 44 .Previous work by others has shown that TH is instrumental in determining the pigmentation patterns present in zebrafish skin and maintaining proper pigmentation patterns in juveniles and adults 33,38 .Ablating the thyroid of larval zebrafish resulted in significant skin pigmentation changes after 6 months, particularly an increased number of melanophores and wider stripes 33 .Additionally, a genetically hyperthyroid mutant (opallus) showed an increased number of xanthophores and decreased number of melanophores 33 .The opallus mutant experiences hyperthyroidy from a very early age.Therefore, we saw the opportunity to test whether the adult pigmentation patterns of euthyroid zebrafish were plastic to more "acute" hyperthyroidy through treatment of adults with T4.We found that WT zebrafish exhibited striking skin pigmentation changes after 5 days of TH treatment.We observed that in T4-treated fish, the dark stripes appeared to lighten and appear green (Fig. 6A,B), and the fish pigment patterns appeared qualitatively similar to the opallus mutant 33 .Using photoshop assisted spectroscopy 45 , there appeared to be trends toward differences in stripe color but not interstripe color (Fig. 6C).This change did not occur in the control group (Fig. 6D).These results suggest that skin pigmentation in zebrafish is plastic to external signals even at the adult stage and in response to only 5 days of exposure, and that specific TH levels are required to maintain homeostasis in skin pigmentation.

Discussion
Cone subtype patterning in the adult zebrafish is plastic.Our results show that exogenous TH treatment alters opsin expression in as little as 7 h.When adult fish were treated with T4 for 24 h or longer, the expression of several genes was affected.Lws1 expression increased while the expression of lws2, gngt2b, rh2-1, and rh2-2 decreased.Further, we found that when adult fish were treated with T4 for 12 h or less, lws1 expression increased but the expression of the other transcripts tested did not change.Additionally, we found that exogenous T4 treatment for 5 days alters skin pigmentation in adult zebrafish.Because the appearance of the T4-treated fish was similar to that of the adult opallus mutants, and the opallus mutants exhibit increased xanthophore numbers and decreased melanophore numbers 33 , it is likely that the T4 treatment also increased xanthophore numbers and decreased melanophore numbers.Our results provide further evidence that homeostatic TH levels are required for zebrafish to maintain normal melanophore and xanthophore populations, and to maintain the opsin expression patterns characteristic of adult zebrafish.
We focused this study upon selected cone photoreceptor transcripts (lws1, lws2, gngt2a, gngt2b, rh2-1, and rh2-2), and found that these remain plastic to TH treatment in the adult zebrafish.However, the plasticity of other transcripts such as rh2-3 and rh2-4 or sws1 and sws2 remains unknown.As T3 levels in a particular cell can be specifically tuned by local deiodinase enzymes, it is possible that cells in various regions of the retina effectively received different amounts of TH 27 .Additionally, data for our 7 h experiment were collected at a different time of day than our other experiments (Supplementary Fig. S4).As zebrafish opsin expression exhibits circadian changes, with opsins expressed at low levels in the morning and high levels in the evening [46][47][48][49] , this represents a confounding variable in our evaluation of the 7 h treatment results.In pigmentation analysis, our methods were limited to photography and Photoshop-assisted spectroscopy.Analyses using pigment cell counts or spectrometry would strengthen these results.
The present study advances the extensive literature examining the role of TH in altering or maintaining cone phenotypes in fish.TH signaling has been shown to be important prior to, or during smoltification, a post-embryonic life stage transition that occurs in salmonids as juvenile fish change habitat 50,51 .This transition is accompanied by changes in skin pigmentation (lightening) and cone photoreceptor subtype patterning (long wavelength-shifting) [52][53][54] .In both coho salmon and rainbow trout, TH signaling is associated with a UV to blue shift in opsin expression that begins toward the end of yolk sac absorption when there is a surge in TH 22,53 .TH signaling also underlies changes in opsin expression during metamorphoses of two flounder species 55,56 .Further, TH has been shown to induce the expression of cyp27c1, an enzyme within the retinal pigmented epithelium that converts vitamin A1 to vitamin A2, and thereby long wavelength-shifting pigment sensitivity 14,25,57 .In coho salmon, a chromophore shift naturally occurs and is associated with seasonal variables 58 .Interestingly, there is some evidence that TH-induced plasticity in photoreceptor gene expression may be conserved in humans.Cakir et al. 59 found that adult patients treated for hypothyroidism showed significant improvement in their color contrast sensitivity, indicating that cone gene expression in the adult human retina may also be plastic to TH levels.The ability of TH to alter gene expression in the photoreceptors of an adult organism could inform therapeutic approaches to disorders involving cone photoreceptors, or in determining optimal protocols for retinal organoid development or other photoreceptor replacement strategies.
Thyroid hormone receptors (thraa, thrab, thrb), deiodinases (dio1, dio2, dio3a, dio3b) and thyroid hormones (T3/T4) are present in the adult zebrafish retina 39,57 ; however, to our knowledge the topographical distributions of these remain unknown.Retinoic acid signaling, however, is known to occur in the ventral portion of the retina in embryonic, larval [60][61][62] , and juvenile (1 month) zebrafish 36 .In thyroid-ablated juvenile lws:PAC(H) reporter fish, lws1:GFP+ cones are only present in a ventral patch that correlates with the location of RA signaling, showing that in the absence of TH, RA signaling is sufficient to promote the expression of lws1 14 .Further studies are needed to examine the topography of both TH and RA signaling and their signaling components in the zebrafish, particularly in adult retinas.
Based on results from our previous studies in larval zebrafish, our data suggest that the cones co-expressing lws1 and lws2 identified in our 12 and 7 h experiments are likely undergoing opsin switching 14 .The results from the present study additionally reveal that the kinetics of T4 treatment and changes in gene expression within adult retinas are complex.Since zebrafish have two loci for each specific opsin gene (all are on autosomes), the coexpression could be achieved either by the expression of both genes on each locus, or by the expression of one gene from one locus and the other gene from the second locus.Alternatively, it is possible that lws2 was not being transcribed and the mRNA from hours before had not yet been degraded.Other factors, including a temporal difference in the mechanisms controlling lws1 versus lws2 expression could also underly this observation.For example, the mechanism for regulating lws1 may involve direct interaction of a liganded TH receptor with elements on the lws1 locus, while control of lws2 may be indirectly mediated and/or require recruitment of additional co-repressors.Interestingly, in larvae, 24 h of treatment with 100 nM T3 produced similar results to the adult 12-h treatment, in which lws1 levels increased while lws2 levels did not change, but by 48 h of larval treatment, lws2 levels decreased 14 .The faster rate of TH-induced changes in adults could be due to the higher concentration of TH in the system water (100 nM for larvae vs. 386 nM for adults), or other factors such as changes in the temporal controls of TH response between larvae and adults.
While the exact nuclear hormone receptors responsible for controlling lws1 vs lws2 gene expression remain unknown, TH receptor beta 2 (thrb2) is a reasonable candidate for this role.Thrb2 is required for LWS cone development in zebrafish 57,63,64 and sequences having predicted TH receptor binding activity have been identified near the lws1 and lws2 genes 14 .Additionally, overexpression of thrb2 in zebrafish cones and bipolar cells has been shown to long wavelength-shift the sensitivity of the adult retina, with the response data fitting a vitamin A1-based model with maximal amplitudes near the measured amplitudes for zebrafish LWS1 cones 13,65 .Adult zebrafish retinas experience several surges of TH during the zebrafish lifespan 33,57,66 , and so the shift from LWS2 www.nature.com/scientificreports/cones to LWS1 cones in the thrb2 overexpression model may represent the consequences of predominantly liganded TH receptor.The complex kinetics of TH-mediated changes in adult photoreceptor gene expression could be mediated by the ability of the Thrb2 receptor (or other nuclear hormone receptor) to function in a manner that differs from the canonical model of nuclear hormone receptor action, in which the presence of ligand leads to recruitment of coactivators and the absence of ligand leads to recruitment of corepressors at genes that are positively regulated by TH 31 .Indeed, it has been shown that both liganded and unliganded forms of Thrb2 can promote transcription of positively regulated genes 67 , although the activity level of liganded receptor is higher.Further, thrb1, a splice variant of thrb2, has been shown to control gene expression by altering ratios of coactivators and corepressors, rather than recruiting either coactivators or corepressors 68 .Roles for TH receptors other than Thrb2 are also possible; for example, TRα has been implicated in UV-to-blue-sensitive opsin switching in salmonids 21 , and during TH-induced metamorphic changes in opsin expression in the winter flounder 56 .
The results shown here emphasize the necessity for the identification and study of the transcription factors that regulate tandemly replicated opsin genes.We observed that both rh2-1 and rh2-2 were downregulated by TH treatment in adults.In contrast, rh2-2 is upregulated by exogenous TH in larvae and unchanged in juveniles 14 .While the expression domains of the lws and rh2 opsins shift through the juvenile stage, the expression domain of rh2-2 is particularly dynamic as the fish grows/ages 34 .In the embryo, rh2-2 is expressed both centrally and peripherally.In the juvenile, rh2-2 is expressed in the dorsal periphery and ventral mid-periphery, and in adults, rh2-2 is expressed centrally 34 .Previous work has shown how cis elements of the rh2 array underly expression domains of the rh2 genes in adult zebrafish 69 .Our results here provide additional insight into how TH may also be involved in tuning the expression of rh2 genes, and that this tuning effect may change over the zebrafish lifespan.
In addition to regulating cone opsin gene expression and the type of opsin chromophore, TH has been shown to regulate multiple photoreceptor transcripts, including multiple transcripts that are differentially expressed in LWS1 and LWS2 cones such as gngt2b 20 .Recent work has shown that zebrafish cone subpopulations exhibit inter-population transcriptional heterogeneity and intra-population heterogeneity, with gngt2b as an example of a transcript that varies in expression between the LWS cone subtypes, and within the cone subtypes 20,26 .Other work has implicated thrb2 in regulating multiple genes that are expressed in spatial gradients in mouse retina 70 .Taken together, these results implicate TH as an important regulator of transcriptional heterogeneity in cone populations.Because the non-opsin, LWS2 cone-enriched gene gngt2b exhibited plasticity to TH treatment in adult fish, it is possible that additional cone-expressed genes also remain sensitive to TH treatment in the adult.Therefore, TH may be involved in regulating transcriptional heterogeneity among and between cone subpopulations in adult zebrafish.
TH signaling is an important regulator of life history transitions in fish 50 and other vertebrates such as frogs 71 .Many of these life stage transitions are accompanied by a change in habitat or ecological niche, requiring different visual system capabilities 52,55,56,71 .Zebrafish also undergo TH-mediated changes in jaw morphology, pigmentation, and feeding strategy as they change from larvae to juveniles 32,72,73 .The plasticity we observed in adults, however, cannot be directly explained as relevant to a life history change, as we used captive, adult, reproductively mature zebrafish.This plasticity could instead indicate that TH signaling serves as an ongoing mechanism for maintaining cone subtype patterning in adults.Indeed, TH gradients in adult mice are involved in maintaining cone subtype patterning 15,24 .As zebrafish possess the ability to regenerate their retinas after injury, the plasticity of adult cones could also be important in re-establishing some elements of cone subtype patterning in the regenerating retina 74,75 .There is evidence that zebrafish re-establish near-normal topographic patterns of lws1 vs lws2 after extensive damage to retinal neurons and subsequent retinal regeneration 75 , and TH signaling may underlie this phenomenon.Plasticity within the visual system of cichlids experimentally exposed to different lighting conditions has been demonstrated, although the underlying mechanism(s) are not known 76 .It is possible that non-captive zebrafish utilize an endocrine mechanism to adjust their visual system to changing environmental conditions.For example, cichlids respond to increases in turbidity by long wavelength-shifting visual function 77 , and larval and juvenile Atlantic halibut respond to white-to-blue light environmental change by increasing density of LWS cones.Further, the overall cone photoreceptor pattern transitions from a hexagonal lattice to a square mosaic, likely through opsin switching 78 .
The present study builds upon our previous work showing that TH regulates the expression of the lws and rh2 opsins in larval and juvenile zebrafish, by determining the extent to which the zebrafish retina is plastic to TH, and reveals an interesting difference in the TH response kinetics of lws1 and other cone photoreceptor genes.We found that skin pigmentation in adult zebrafish also remains plastic to exogenous TH treatment, showing an overall plasticity to TH that is reminiscent of TH-mediated postlarval photoreceptor and pigmentation changes seen in salmonids.This work adds to the body of literature showing TH is an important regulator of retinal development and cone subtype patterning, as well as an essential driver of retina and pigment phenotype changes in fish.

Ethics statement
All animal experiments were performed in accordance with relevant guidelines and regulations and with approval from the University of Idaho Institutional Animal Care and Use Committee.Methods are also reported in accordance with ARRIVE guidelines.

Animals
Zebrafish were propagated and maintained according to Westerfield, on recirculating, monitored, and filtered system water, on a 14:10 light/dark cycle, at 28.5 °C35 .Procedures involving animals were approved by the Animal Care and Use Committee of the University of Idaho.Wild-type (WT) zebrafish were of a strain originally ] harbors a PAC clone that encompasses the lws locus, modified such that a GFP-polyA sequence, inserted after the lws1 promoter, reports expression of lws1, and an RFP (dsRedExpress)-polyA sequence, inserted after the lws2 promoter, reports expression of lws2 40 .This line was the kind gift of Shoji Kawamura and the RIKEN international resource facility.Adult (0.5-1.5 years) zebrafish were used.Both sexes were represented in each control and treatment group for all experimental endpoints.Zebrafish were considered adult if they were over 6 months of age, showed adult pigmentation patterns, and displayed adult male/female body shape characteristics (slim for males, plump for females) 35 .

Thyroid Hormone Treatments
Stock solutions of tetra-iodothyronine (T4; Sigma) were prepared in NaOH (Sigma), and maintained at − 20 °C in the dark.During treatments, adult zebrafish were maintained individually in 250 mL beakers in system water.10,000× T4 stock solution was added to system water for a final concentration of 386 nM as in Suliman et al. 14,38 (NaOH final concentration was 0.01% and did not alter system water pH).Controls were treated with 0.01% NaOH.For experiments lasting > 1 day, fish were fed once daily and treatment solution was completely replaced after feeding.Duration of treatments was 7 h, 12 h, 24 h, or 5 days (Supplementary Fig. S4) 14 .T4 (rather than T3 or a synthetic analog) was used as the experimental treatment to be consistent with other studies of TH treatment in postlarval fish 14,25,33,79 , and with the rationale that TH would primarily enter via the gills into the bloodstream, and T4 is the predominant, circulating form of TH 28 .Blinding investigators to treatment condition was not possible due to the need for proper solution changes, overall appearance of the fish for the 5 day treatments (Fig. 6A), and the large increases in lws1 expression in the T4 treatment groups for all treatment durations (Figs.

RNA extraction and quantitative RT-PCR (qPCR)
Total RNA from zebrafish eyes was extracted using the Machery-Nagel Nucleospin RNA kit, and then the Superscript III/IV (Invitrogen) was used to synthesize cDNA template with random primers.Gene-specific primer pairs for qPCR are provided in Supplemental Table S1.Amplification was performed on a StepOne Real-Time PCR system using SYBR Green or Power Track SYBR Green master mix (Applied Biosystems).Quantification of transcript abundance was relative to the reference transcript (β-actin), using the ddCT method.Transcript abundance for β-actin did not vary significantly between control and treated groups.Graphing and statistics were performed in Excel.Sample groups were evaluated for normal distributions using the Shapiro-Wilk test.

Histological processing
Fixation and preparation of adult PAC(H) eyes for tissue sectioning were performed as previously described 14,36,74,75 .In brief, zebrafish were humanely euthanized (anesthetized using MS222 overdose followed by decapitation), eyes were removed, corneas pierced and lenses removed.Eyecups were then fixed with 4% paraformaldehyde in a phosphate-buffered, 5% sucrose solution overnight at 4 °C, washed in increasing concentrations of sucrose, cryoprotected overnight at 4 °C in phosphate-buffered 20% sucrose, embedded and frozen in a 2:1 solution of 20% sucrose: OCT medium (Sakura Finetek, Torrance, CA), and sectioned at 5 μm thick 60 on a Leica CM4050 cryostat.

Hybridization chain reaction (HCR) in situ hybridization
HCR procedures were carried out according to the manufacturer's instructions (Molecular Instruments), with the exception that we did not incorporate a proteinase K treatment prior to the post-fixation step.In brief, zebrafish retinas were dissected and fixed overnight in 4% paraformaldehyde in phosphate-buffered saline (PBS) at 4 °C.Tissues were then washed in PBS, dehydrated in MeOH, and stored in MeOH at − 20 °C at least overnight.Tissues were rehydrated in a graded MeOH/PBS/0.1% Tween 20 series, and post-fixed with 4% paraformaldehyde in PBS prior to hybridization.Hybridization was done overnight at 37 °C.Tissues were washed with the manufacturer's wash buffer, and then 5XSSCT (standard sodium citrate with 0.1% Tween-20), and the amplification/ chain reaction steps were performed following the manufacturer's protocol.Transcript-selective probe sets were designed and generated by Molecular Instruments (Supplementary Table S2) and can be ordered directly from their website.

Confocal microscopy
Whole, fixed lws:PAC(H) adult retinas, cryosections of lws:PAC(H) eyecups, and HCR-processed, adult (0.5-1.5 years) WT retinas were mounted in glycerol and imaged with a 20× dry lens using a Nikon-Andor spinning disk confocal microscope and Zyla sCMOS camera running Nikon Elements software (RRID:SCR_014329), and 3 µm-step sizes were used for Z-series images.Z-stacks were flattened by max projection, and brightness/ contrast adjusted in FIJI (ImageJ) (RRID:SCR_002285).Multiple images encompassing the entirety of whole retinas or whole cryosections were stitched together using the "large stitched image" feature in Nikon Elements Software. Vol:.(1234567890)

Analysis of comparative areas of expression domains
Expression domains were traced using the freehand measurement tool in FIJI/ImageJ as described in Stenkamp et al. 75 .Areas containing predominantly or exclusively GFP+ cones or RFP+ cones were measured within the individual fluorescence channels while areas containing "interspersed GFP+ and RFP+" cones (including areas containing co-labeled cones) were measured using both the red and green channels.Each area was traced in triplicate with the freehand selection tool in FIJI to ensure measurement reproducibility.Percentages were determined by dividing the number of pixels in each expression domain by the number of pixels in the entire retina 75 .This strategy was also applied to measure domains of native lws1 and lws2 expression in WT whole retinas imaged after HCR in situs.The Shapiro-Wilk test was used to check for normal distribution.Differences between each measurement were tested using Student's t-test.Overall difference between control and treatment was tested using Fisher's Exact Test.

Brightfield photography and color measurements
Fish were anesthetized using MS-222 and placed on a sterilized portion of the lab bench.Fish were photographed using a Canon PowerShot SX70 HS, and each fish was photographed with and without flash, with a focal length between 40 and 60 mm.Stripe and interstripe colors of individual zebrafish were measured using the RGB spectrum tool in Adobe Photoshop (RRID:SCR_014199) 45 .

Figure 6 .
Figure 6.Five day T4 treatment of adult zebrafish alters skin pigmentation.(A) Brightfield photographs of representative fish.(B) Examples of representative stripe and interstripe colors of control and treated fish.Colors were sampled from regions denoted by rectangles in (A).(C) Color analysis of stripe and interstripe Red, Green, and Blue scale values, n = 3.Values denote "redness, " "greenness, " and "blueness" on a scale from 0 to 255.High red values indicate high redness.Low red values indicate green, low green values indicate red, low blue values indicate orange.Note trend of higher red and green values in stripes of T4-treated fish.(D) Percentages of fish in control and treated groups with altered pigmentation compared to untreated wildtype.All T4-treated fish showed altered skin pigmentation while all control fish showed normal pigmentation, p = 0.0143 (proportion test).